A Prediction of breast cancer based on Mayfly Optimized CNN
T. Gayathri, T. Madhavi, K.Ratna Kumari · 2022
Breast cancer patients are prone to major problems of greater mortality related to their health. Breast cancer diagnosis is time demanding, and a system that can automatically diagnose breast cancer in its early stage has to be developed due to the low availability of systems. The principal explanation may be that radiologists are misinterpreting suspected injuries due to technical difficulties in picture properties and diverse breast densities, hence increasing the false ratio. Timely intervention is important in developing an upto-date forecasting method that can reduce disease complications with increased recovery. For the categorization of benign and malignant tumours, various Machine Learning (ML) and Deep Learning (DL) algorithms were applied. However the detection by typical machine learning methods of breast abnormalities misinterprets an inconsistent feature extraction process that causes problems, the call-back to patients for biopsies to eradicate suspicions. Deep learning approaches for a reliable pronostic and categorization of breast cancer have therefore been developed. This research study focuses on the prediction of breast cancer using Convolutional Neural Network (CNN). The input images are removed by noises and segmentation is carried out by Otsu Thresholding technique. After that, CNNis applied for final prediction, where the learning rate of the CNN is optimized by using Mayfly Algorithm. The experiments are validated to test the efficiency of proposed CNN and proved that it achieved 97% of accuracy for the publicly available dataset.